jablonkagroup/ChemBench-Results-dev

这是一个包含模型信息的的数据集，其中包括模型的ID、名称、参数数量、详细描述、发布日期、相关论文和代码链接等。数据集还提供了模型是否开放权重、是否为混合专家模型、模型对齐方式、是否通过人类反馈进行强化学习等信息。此外，数据集还记录了模型在多个化学领域的评分。数据集分为两个版本，1.0.0和1.0.1，分别包含32和33个示例。

This dataset contains information about models, including model ID, name, number of parameters, detailed description, publication date, related paper and code links, etc. It also provides information on whether the model opens weights, whether it is a mixture of expert models, model alignment, and whether it learns reinforcement from human feedback. In addition, the dataset records the scores of the model in multiple chemical fields. The dataset is divided into two versions, 1.0.0 and 1.0.1, containing 32 and 33 examples respectively.